Asynchronous vs Multithreading and Multiprocessing Programming (The Main Difference)

what is going on guys my name is russain and in this video I want to talk about three four four things yes synchronized operations asynchronous operations multi-threading and multi processing I hope I nail all of them if you are interested stay tuned so the first one guys synchronous operation synchronicity is when you make a request whether this is a network call or a request to read from disk or a request to send to a printer to print does anybody print these days I don't think so right but yeah if you're doing that then there is this magic thing that's what we call the process and there's in this process there's always by default one thread that does the work actually does the actual work now think about it guys if you're calculating a prime number then you that thread is just busy in the CPU right this is using the resource of the CPU to calculate things 1 plus 1 divided by 3 how do you calculate a prime number alright so now it's busy but think about it if you're reading from disk you ask your thread is asking the operating system to read a file from disk the thread is not doing actually any work it's basically blocked and waiting that's it it's just waiting so if you reading then who who is doing the actual work from reading the disk that's the disk controller it's called IO controller it's it's something on the motherboard that I forgot how it actually works it's just some driver on the battle board maybe the operating system is responsible that so the thread here is just sitting here like that it's just yeah alright I'm just waiting here not doing nothing so the rest of your code guess what it's also blocked and that's called synchronous execution because you're sending a request and you're waiting and blocked despite you not doing any work really is just blocked so that's synchronous exclusion same thing applies to ascending a rest request to the endpoint to the to the server right you're not doing any work the server is doing the work you're just waiting for result right the TCP just file descriptor is waiting for something to wake it up and receive them for mayhem well until that is you also block so if there are code here you cannot execute that code unfortunately if there is a UI thread right no if you if there is a user experience or UI and peep the user is actually pressing assuming you're a single threaded that button doesn't do anything you cannot do anything because that thread cannot leave that state to actually go and serve and animate the button as if it was clicked because guess what it's blocked assuming a single threaded right most application has a UI thread in but let's assume just for simplicity if you're asking a visual basic 5 application like 1998 kind of a things right but yeah one application is blocked right now people say this is a ridiculous we cannot do this this is dumb how about we introduce some other thread let the let the process execute another thread that does more work sounds cool so what is a thread a thread is we have the process think of it as like a container but think of it like almost like a container that has 1/3 of execution all the time right if you as the process owner a writer of the application can spin up another threat to do some other work another third can do some other work all these threads have in common is they share the same resources that the process has that is big big big big big problem because they can start tracing for these memory locations and you can get into the same situation that you get with a database which is locking and race conditions and all that jazz right so what do you do guys what do you do right you don't do multi-threading because multi-threading is evil right check out this white paper that people just say stay away from multi-threading so people says okay some people stayed away from multi-threading because they cannot get it right because it's very hard to build your application so it's a thread safe right that's what calls thread safety is essentially having mutexes on your resources so no two threads can actually access the same process resources at the same time what's the problem right that's the that's the big problem right so people shied away from multi-threading some people like there's still a thing some people still doing here and they are bragging that their applications through itself definitely sequel light per hour cycle light does not like multi-threading check out check that out he despised it so yeah so multi-threading is essentially evil that's my personal opinion you might have different let me know in the comment section but essentially multi-threading is yeah it's very hard to write it's very very difficult to get right you need to have a very very skillful understanding of how operating system works and your CPU because your CPU can have multi cores and your threads can execute in parallel literally important not time slice in parallel like that's that's good but doesn't give you much in for much benefit to be honest some smart people came and says you know what we're gonna execute something called asynchronous accusin and nodejs is a very elegant example of this it's a single threaded non blocking asynchronous framework that's that's the title of this so let's talk about asynchronous ctz right i cannot spell guys and i cannot talk about asynchronous cities asynchronicity is the idea of having a single threaded single process single thread that does the work the process cannot do work by itself well it needs a thread but well that is also based on the operating system Linux treat threads as a sub process versus windows actually just three stories either actual thread of thing which would kind of complicate things but never Maya back to the point single thread single process if you're reading from desk we're gonna do it asynchronously here's what I'm gonna do because the thread doesn't literally it is not doing anything it's just waiting and it's being blocked and the rest of the code cannot be executed it's blocked here right and the rest of the code can't get executed right so what do we do what do we do guys and I'm sending a request to the i/o controller to read from disk and I'm sending a request a fetch command for example JavaScript and I'm reading from network resources breast endpoint G RPC right and I'm waiting I'm just just sitting here waiting doing nothing right the application knows this and you say oh okay this thread is not really doing anything so what we're gonna do is let's send that request that is essentially that a quiz that is actually was blocking and here's what I'm gonna do we're gonna tell whoever owns that other end io control network says hey here's a request I want to read this resource I want to read this rest endpoint I want to read this block of desc and here's the thing I'm gonna go away and here's a function it's called a callback call me back all right whenever you're done I'm gonna do my own thing here so the the thread immediately unblocked and start executing other codes until that function finishes right that actual reading is done and the reader will call back that function which will trigger some sort of an event which the thread oh there's something here let me go and execute and this is in JavaScript the implementation is that literally an event loop the thread goes is okay do I have jobs to execute do I have actual code that I need to execute is there something coming from a callback is there something going a lot yeah so that's how work async honestly beautiful beautiful design one slight problem and I don't know if you feel coded and callbacks in Ajax back in the 2000s 2004 2003 that's when I started coding with Ajax that cord is ugly as f you cannot read it because most of the time we need go a callback you need some information from the callback to do something else so you start doing ok callback and then is okay this function and then you take this function and then let me go do all this other thing and then that thing needs a callback so you end up doing a curly braces in the cooler braces in color braces and color braces and it's completely unreadable right so people when they invented this meanwhile some people didn't like it the multi-threaded people says yeah that's just ugly code like multi so it is a beautiful code huh right but some people defer to from this because it's just a cleanest the JavaScript community came in and says you know what we'll fix this we still like I synchronous but we're gonna bring promises they promises and the picture and now the code is better with that vendor then the then the then the catch some people even did it even better and that's pretty much a standard I think async await you write the function you code this is asynchronous function and every code that is essentially blocking you write and await on top of it and that's it it will act asynchronously but look synchronous to you and it's all sheesh and syntactical sugar shin tack she can you think syntactical sugar three times in a row I dare you guys you cannot syntactical sugar I cannot think damn it okay all right go back all right so that's that's a a syncretic sujin that's my favorite thing nor I like my process to have single threaded and I don't care about multi-threaded I'm gonna use asynchronous execution all this mess of synchronous asynchronous and multi-threaded some people said if all that we are multi processing guys multi processing I'm also with this team multi processing and I'm gonna define it the way I understand it if if that doesn't make sense let me know or if he if you disagree let me know because I would love to learn how do you perceive this so multi processing the idea of spin up instead of splitting up threads in a single process which shares the same resource of the process right what the heck is that okay so instead of spinning up multiple threads in a process spit up just unique processes with their own beautiful memory structure with their own resource with their own everything and then what do you do you just communicate between them use inter process communication you can use a centralized Redis database and yeah there are many ways to communicate between processes you can use sockets between them I know it's a little bit overkill but you can list literally listen on a server here localhost localhost localhost and ports and communicate with port and all the communication will happen with ccp's right that's fine it's all localhost anyway who cares you can do that people do this absolutely fine right there are some limitations but here's one example where multi process is actually good let's say you want write a function that brute-force a password that example I know but you want to write it so that multi-threaded application multiprocessors and asynchronous write the best way to write it in my opinion is use a multi processor because multi process can be spin up on multiple not just judgin the same machine could be scaled to run on multiple machines as well alright and i believe this is essentially how containers work in a nutshell they're more isolated than a process but sure but here's a here is a good problem I want - this is a hash right md5 hash and I want to know I want to reverse it which is an almost an impossible problem impossible it is a very hard problem to solve so what do you do well you basically come up with garbage strings and try to hash it and see if they match it and the string matches the hash of this thing matches the hash that you were given that means that's an input is a candidate input right well one machine can just take a string and then just loop through and take a dictionary a rainbow table and just loop through all of them and then generally start generating the hash that's absolutely fun but here's the thing you can have multiple processes throw multiple processes at the problem and you say okay you process one take this rainbow table you process to take this dictionary from A to M and this process take from n to Z right and so on you can stop splitting a problem and multiple process so your multi process sink right and if you're kind of hip right and it's like yeah if you're hip you can use a reverse proxy send that problem to reverse proxy and let the reverse proxy actually divide the problem you can write logic in the reverse proxy I believe it's a proxy Ingenix allow you to write some checks at least lower code to actually split Hey oh you yours this server spin up a container or multiple containers and throw some problems at it it's like between this and this and you server between this and this and you between this and this and so you're kinda multi-processing twindom hadoop kinda work the same way as well but yeah so that's essentially a multi processing that's in the way I see it might be wrong but that one's I can brief discussion between synchronous asynchronous multi-threading and multi processing guys I hope you enjoyed this video you guys see you in the next one keep up those good good questions coming guys I love them so much get see in the next one you guys stay safe out there

Channel: Hussein Nasser

Views: 55,863

Rating: 4.8585114 out of 5

Keywords: sync vs async, multithreading vs multiprocessing, multithreading async, threading, thread-safe, thread safty

Id: 0vFgKr5bjWI

Length: 15min 32sec (932 seconds)

Published: Fri Apr 17 2020